Demand and conversion
Bounce rate and engagement rate
Two mirror metrics for whether a visit went anywhere — and the definition changed underneath everyone when analytics moved to GA4.
Formula
Engagement rate = Engaged sessions ÷ Sessions; Bounce rate = 1 − Engagement rate
- Engaged session
- in GA4, a session lasting over ten seconds, or firing a conversion, or reaching a second page view — any one of the three is enough
- Sessions
- all sessions in the period, on the same property and the same filter
Before GA4 a bounce was a session with exactly one interaction, so the two definitions are not the same quantity and a figure that crosses the migration compares different things. The engagement threshold is configurable, which means two properties can report different bounce rates for identical behaviour.
In the older model a bounce was a session with exactly one interaction, which meant a visitor who read a page for four minutes and left satisfied was counted identically to one who arrived by mistake. In GA4 the primary metric is engagement: a session counts as engaged if it lasts over ten seconds, fires a conversion, or has two or more page views. Bounce rate is simply its complement.
The practical consequence is that bounce figures from before and after the migration are not comparable, and a reported improvement across that boundary is usually the definition changing rather than the site.
On a page that answers one question — opening hours, a phone number, a definition — a visitor who leaves immediately was served perfectly. Judge single-answer pages on whether the answer was found, not on whether a second page was loaded.
Quantities derived from the counts. Each is only as sound as the definitions beneath it.
Cannot be computed without
- Session, user, event
- The three units analytics counts in, each measuring something different and none of them a person.
A definition is free. Being answerable for the figure it produces is the part that is bought, and this term is a working part of the engagements below.
- Intelligence and decision layers
- A metric with two definitions has none.
The definitions are the easy part. Whether the figure on your dashboard was computed this way is a different question, and usually the more expensive one.